{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.01751", "kind": "paper", "url": "https://arxiv.org/abs/2607.01751"}, {"id": "hf:venn2024/medstreambench", "kind": "dataset", "url": "https://huggingface.co/datasets/Venn2024/MedStreamBench"}], "authors": ["Yuan Wang", "Shujian Gao", "Songtao Jiang", "Zhengyu Hu", "Zuozhu Liu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.01751", "source_url": "https://arxiv.org/abs/2607.01751"}, "categories": ["Biology & Drug Discovery", "Vision & 3D", "cs.CV"], "description": {"en": "MedStreamBench is a time-aware benchmark for medical video understanding, integrating 22 medical datasets and 5,419 QA instances across four temporal settings: retrospective, present, future, and proactive. Models are restricted to temporally bounded evidence windows and evaluated on answer correctness, responsiveness, and post-evidence stability."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.01751", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.01751", "title": "MedStreamBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Zhejiang University", "Shanghai Key Laboratory of Trustworthy Computing", "Carnegie Mellon University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.01751", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.01751", "source_url": "https://openalex.org/W7167272475"}, "key": "claire-radar:2607.01751", "modality": null, "name": "MedStreamBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_medstreambench_73e2ea6c", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "a46f47dd004e292bb0b967a642bee0d57d61ca8af735b1a2b4ab4cc8261a3e21", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.01751"}, "publisher": null, "released": "2026-07-02", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.01751", "source_url": "https://arxiv.org/abs/2607.01751"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-01751", "source": "claire_radar", "source_benchmark_id": "2607.01751", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Biology & Drug Discovery"], "area": "Vision & 3D", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": "Venn2024/MedStreamBench", "hfDatasetDownloads": 171, "hfDatasetLikes": 4, "hfPaperUpvotes": 2, "hfPaperUrl": "https://huggingface.co/papers/2607.01751", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-04", "state": "stale"}, "hfDatasetLikes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-04", "state": "stale"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-10-04", "state": "stale"}}, "status": "stale"}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "2d8a770354dccdbddfe653a361c7d0b86c6469f65b5308cb13be3d45b315e70d", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "MedStreamBench is a time-aware benchmark for medical video understanding, integrating 22 medical datasets and 5,419 QA instances across four temporal settings: retrospective, present, future, and proactive. Models are restricted to temporally bounded evidence windows and evaluated on answer correctness, responsiveness, and post-evidence stability.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We present MedStreamBench, a benchmark for time-aware medical video understanding."}, "familyId": "bmf_b4144a83ff7d", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_medstreambench_73e2ea6c", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Pharma & Biotech"], "links": {"code": null, "data": "https://huggingface.co/datasets/Venn2024/MedStreamBench", "hfPaper": "https://huggingface.co/papers/2607.01751", "paper": "https://arxiv.org/abs/2607.01751", "pdf": "https://arxiv.org/pdf/2607.01751", "project": null, "report": "https://arxiv.org/abs/2607.01751"}, "metrics": [], "motivation": "Existing medical video benchmarks primarily evaluate whether a model produces the correct answer, but rarely assess whether it answers at the right time.", "name": "MedStreamBench", "oneLine": "MedStreamBench is a time-aware benchmark for medical video understanding, integrating 22 medical datasets and 5,419 QA instances across four temporal settings: retrospective, present, future, and proactive. Models are restricted to temporally bounded evidence windows and evaluated on answer correctness, responsiveness, and post-evidence stability.", "paperTitle": "MedStreamBench: A Time-Aware Benchmark for Streaming and Proactive Medical Video Understanding", "primaryDomain": "Biology & Drug Discovery", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.01751", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-02", "source": {"authors": ["Yuan Wang", "Shujian Gao", "Songtao Jiang", "Zhengyu Hu", "and Zuozhu Liu"], "categories": ["cs.CV", "cs.AI"], "id": "2607.01751", "title": "MedStreamBench: A Time-Aware Benchmark for Streaming and Proactive Medical Video Understanding", "type": "arxiv", "url": "https://arxiv.org/abs/2607.01751"}, "sourceUpdatedAt": "2026-07-02T06:07:44Z", "topics": ["cs.CV"], "venueAttempts": [], "whyItMatters": "It addresses the gap between offline recognition and temporally grounded decision-making in clinical settings, where models must decide when to answer or alert. The benchmark provides a protocol for evaluating time-aware reasoning in medical video."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
